Skip to Main Content
 

Global Search Box

 
 
 
 

Files

File List


case1212774902.pdf (1.15 MB)

ETD Abstract Container

Abstract Header

Improving Genetic Analysis of Case-Control Studies

Won, Sungho
http://rave.ohiolink.edu/etdc/view?acc_num=case1212774902

Abstract Details

2008, Doctor of Philosophy, Case Western Reserve University, Epidemiology and Biostatistics.
The improvement of genotyping technology makes genetic association analysis feasible at the genome-wide level. However, the computational burden for multi-locus analysis is still hard to handle and often only single SNP analysis with the Cochran-Armitage test is used. For testing association, the possible evidence comprises three typesof information: differences between cases and controls in allele frequencies, in parameters for Hardy Weinberg disequilibrium (HWD) and in parameters for linkage disequilibrium (LD) of markers. Also, several SNPs near the causal gene may be associated with disease. Thus, an approach that is computationally efficient and uses all possible information is necessary for inference of association. Biologically, we expect that, at the causal loci, affected individuals are more related evolutionarily than a random pair of individuals from the population and it is known that LD between marker and disease alleles is generally inversely proportional to the distance between them. However, real data show that there can be non-informative markers even near the causal SNP, which indicates that multi-locus analyses based on haplotypes should be considered for locus estimation. Here, first we propose the optimal method for combining p-values using numerical integration or a Monte-Carlo algorithm. The proposed method is always most powerful under certain specified conditions and this is confirmed by simulation. With this method, we propose combining either the p-values from three types of information in a single SNP or the p-values from the Cochran Armitage tests of two consecutive SNP markers. Our simulation and application to the Wellcome Trust data show that the proposed method improves statistical power. For locus estimation, while the coalescentbased approach is biologically reasonable but requires a large number of parameters to be estimated, the clustering-based approach is computationally efficient but biologically less reasonable. Thus we propose for fine-scale mapping a haplotype-based clustering algorithm that is constructed through a Bayesian partition model with a generalized similarity measure, and then extend this method to handle phase-unknown haplotypes. Our simulations show that the accuracy of the estimated location of a causal SNP can be improved by using our new algorithm.
Robert Elston (Committee Chair)
Yuqun Luo (Committee Co-Chair)
Zhu Xiaofeng (Committee Member)
Sunil Rao (Committee Member)
Jing Li (Committee Member)
180 p.
Biostatistics

Recommended Citations

Citations

  • Won, S. (2008). Improving Genetic Analysis of Case-Control Studies [Doctoral dissertation, Case Western Reserve University]. OhioLINK Electronic Theses and Dissertations Center. http://rave.ohiolink.edu/etdc/view?acc_num=case1212774902

    APA Style (7th edition)

  • Won, Sungho. Improving Genetic Analysis of Case-Control Studies. 2008. Case Western Reserve University, Doctoral dissertation. OhioLINK Electronic Theses and Dissertations Center, http://rave.ohiolink.edu/etdc/view?acc_num=case1212774902.

    MLA Style (8th edition)

  • Won, Sungho. "Improving Genetic Analysis of Case-Control Studies." Doctoral dissertation, Case Western Reserve University, 2008. http://rave.ohiolink.edu/etdc/view?acc_num=case1212774902

    Chicago Manual of Style (17th edition)

case1212774902
896
© 2008, all rights reserved.
This open access ETD is published by Case Western Reserve University School of Graduate Studies and OhioLINK.